Hi friends, before we dive in, two things.
First, if you haven’t pre-ordered Crypto Confidential, you should! Morgan Housel said “every investor should read this book,” and Kyla Scanlon said it was “a soulful and a beautiful exploration of an industry that should ask more of itself.” I know you’re going to love it.
Second, if anyone you know is having a baby soon, tell them to check out Gather. They’re providing education & community for postpartum women featuring interviews & chats with experts like Emily Oster. Not an ad, I just know the team working on it and think it’s an important mission. Postpartum resources are woefully difficult to find in a digestible form.
One of my favorite essays is Wendell Berry's "Why I Won't Use a Computer."
In the essay, Berry explains that he prefers to do all of his writing long hand or on a typewriter because he's not convinced that computers have made anyone's writing any better.
And he's right. Going from a typewriter to a word processor does not enhance your writing in any way.
You might be able to do it a little faster or edit a little bit easier, but is the writing going to get any better? Probably not.
If you look at all of the books that were published in the era of long hand and then the era of typewriting and then the more recent era of computing, you couldn't say that the books published now are any better. There are more, but the quality hasn't necessarily improved thanks to the technology. Technology didn’t make you any better.
And while that’s been true for the history of technology and writing, I think it’s changing now.
I've been experimenting with AI tools since GPT-3 became available to the public. When they first came online almost two years ago, the results were okay, but it was more of a parlor trick. It couldn't really produce anything useful and interesting to read, and if you asked it to improve your writing, it couldn’t do much for you.
But that's all changed now with sufficiently good fine-tuning and prompt engineering. You can get Claude III Opus to do a remarkably good job matching your style and tone, expanding on your ideas, and editing your work. It's actually at the point where it can replace the work that many editors would do. And just as junior developers are losing work to AI tools, junior copywriters and ghostwriters probably will as well.
Some people are starting to toot the horn that AI has failed to deliver on its promise. That it’s just a toy. But they’re wrong. What they’re really telling you is that they lack the skills and imagination to use the tools to their fullest. They're like a novice driver stuck in neutral, engine revving but going nowhere, oblivious to the incredible machine at their fingertips.
Two years ago, I was a lot more skeptical. I felt like, "Yeah, there might be something here, there might not be, but we'll see if anything comes out of it.”
But in the last couple of months I've really come around to the idea that no, there is something extremely useful here and you should take the time to try to learn how to use it and how you can use it to make and how to augment your own work.
Given how quickly these tools are improving and how useful they're becoming, the only way that you still have a job in a creative field like writing in the future is if you start learning how to use the tools now.
But I’m not suggesting you should be scared; rather you should be excited. This is one of the greatest opportunities of your life to create a competitive advantage for yourself.
Most writers are not super technical, most artists are not super technical, most editors are not super technical. If you can learn how to work with AI now, you have this window of opportunity to get way ahead of the other people in your field in a way that new technologies haven't let you do to this extent before.
Berry's criticism of computer writing was correct, a computer did not make you a better writer, but AI certainly can. Not by letting it do the work for you but by using it the way you would hire a professional editor to look at everything you work on.
I'll give one last example because I think it's a powerful one.
When computers first started figuring out the game of chess, there were three very distinct phases.
The first phase was that the computers would reliably lose to the humans. They couldn't beat them at chess.
Then, there was a phase where the computers could sometimes beat the humans at chess, but not always. Humans still stood a chance. During that period, the people who won the most were actually human-computer teams. A human could sometimes beat a computer, a computer could sometimes beat a human, but a human and computer working together would have the highest chance of winning.
But eventually, that phase ended, and the computers dominated everyone.
Maybe one day we'll get there with these creative things, too. Maybe one day, writing will actually get worse if there's any human involved.
But we’re not there yet. We're in this sweet spot, like during that period with chess, where you have an incredible opportunity to make the most of it.
Don't get stuck in neutral.
If you enjoyed this, be sure to subscribe so my next essay goes straight to your inbox.
Since I know the immediate follow-up question from this post is going to be, “What do I do?” My best answer is that it’s like learning to use any tool, you just have to start.
But second to that, follow Dan Shipper’s work. He’s providing a ton of fantastic use cases and examples of how people are starting to use it. He might give you some ideas for how to get started.
I was expecting you would end this article with something like "Oh, by the way, this was writen by my AI assistant." So.. not much impressed
Well said. I was also skeptical a couple years ago. My biggest moments of ‘this is a something burger’ was GPT 4 analytics and Claude III is blowing my mind.
The thing to keep in mind is that these tools will probably kinda suck in hindsight a year from now which is wild.
I use these tools every day in my workflow and the compounding I’ve seen is amazing